You are on page 1of 1

General definitions

Spanning tree
Needs to have all vertices of original graph. Delete an edge
Induced subgraph
Delete a vertex and its adjacent edges

Family of graphs
W n Wheel Graph
vertices

edges:

2(n 1)

diameter:

2 n>4

girth:

(Kn ) chromatic number:

3 if n is odd
4 if n is even

W d (k, n) where n number of


Kk joined at a shared vertex

Windmill
graph

vertices

(k1)n+1

edges

radius

diameter

girth

3 if k>2

(Kn ) chromatic number:

k(k1)
2

Kn complete graph
vertices

degrees at each vertex

n1 regular

edges:

n(n1)
2

1n

radius:
diameter:
girth:

(Kn ) chromatic number:

2
1n 2
3n > 3
n

Km,n complete bipartite


graph

vertices

n+m

edges:

n*m

radius:

2 m!=1 or
n!=1

diameter:

2 m!=1 and
n!=1

girth:

4 m!=1 or
n!=1

(Kn ) chromatic number:

Cn Cycle graph
vertices

degrees at each vertex

2 regular

edges:

girth:

(Kn ) chromatic number:

3 if n is odd
2 if n is even

Special graphs
peterson graph
cages

1.1 Graphs and Degrees of Vertices


A graph G is a pair of sets (V, E) where V is nonempty, and E is a (possibly empty) set of unordered pairs of
elements of V.
V(G) - set of vertices
E(G) - set of edges
p = |V| # of vertices
q = |E| # of edges
simple graph
no multiple edges or loops
multigraph
allows multiple edges between 2 vertices, but no loops
pseudograph
allows multiple edges between 2 vertices and loops (edge goes from v1 to v1 itself)
Describing Vertices
Incident
vertex x is an endpoint of an edge e
Adjacent
x and y are vertices of a graph connected by an edge. x and y are adjacent.
Denote the edge xy

Handshaking theorem
Theorem 1.1.1. Let v1 , v2 , . . . , v p be the vertices of a graph G, and let d1 , d 2 , . . . , d p be the degrees
of the vertices, respectively. Let q be the number of edges of G. Then

d 1 + d 2 +. . . +d p = 2q
Theorem 1.1.1 could be stated as follows.
p
p
i=1 d i = 2q or i=1 deg(v) = 2q

This means that the sum of degrees have to be even.


Theorem 1.1.1 (Havel, Hakimi)
How to tell if a sequence is graphic.

1.2 Subgraphs, Isomorphic Graphs


subgraph
of a graph G is a graph H such that every vertex in H is a vertex of G and every edge of H is an edge of G

V(H) V(G) and E(H) E(G)


If two graphs are isomorphic, then they must have the same degree sequence.
Thus, if we are asked to determine if two graphs are isomorphic
1. check whether they have the same degree sequence.

1.3 Trees
Theorem 1.3.1. If G is a connected graph with p vertices and q edges, then p q+l.
Theorem 1.3.2. If G is a tree with p vertices and q edges, then p = q+1.
Theorem 1.3.3. If G is connected, and p = q+1, then G is a tree.
Theorem 1.3.4. Every tree with at least one edge has at least two end vertices.
Theorem 1.3.5. A graph G is a tree if and only if there exists exactly one path between any two vertices
Theorem 1.3.6. Every connected graph G contains a spanning tree.

Most important. A tree cannot have any cycles.

2.1 Vertex coloring


Theorem 2.1.1. Every critical graph is connected.
Theorem 2.1.2. Every graph G contains a critical subgraph H such that (H) = (G).
Theorem 2.1.3. If G is critical with chromatic number four, then the degree of each vertex is at least
three.
Theorem 2.1.3. If G is critical with chromatic number , then the degree of each vertex is at least 1.
Theorem 2.1.4. If G is a critical graph with p vertices and q edges, and G has chromatic number , then
the relation
(1)P 2q holds
Theorem 2.1.6. A graph G is bipartite if and only if every cycle in G has even length.

Let G be a graph.
If H is a subgraph of G, and H G, then H is called a proper subgraph of G.
If (H) (G) for every proper subgraph H of G, then we say that G is critical.
What is a critical graph?
A vertex or an edge is a critical element of a graph G if its deletion would decrease the chromatic number of
G. Obviously such decrement can be no more than 1 in a graph.
A critical graph is a graph in which every vertex or edge is a critical element.
A k-critical graph is a critical graph with chromatic number k;
a graph G with chromatic number k is k-vertex-critical if each of its vertices is a critical element.
X(H) <= X(G) for every proper subgraph H of G
girth
of a graph is the length of the shortest cycle
diameter
The diameter of a graph G is the maximum distance between any two vertices of G.
The diameter of a graph is the length of the shortest path between the most distanced nodes.

proper coloring

Kn is critical
Cn with n odd is critical

2.2 Edge coloring


Describing edges
adjacent
two edges said to be adjacent when they are both incident with the same vertex
A proper edge coloring
is an edge coloring with the additional property that no two adjacent edges receive the same color.
snarks
Those 3-regular graphs that have edge chromatic number four are so interesting that they have a special name
Peterson graph is an example of a snark

Theorem 2.2.1. Let G be a graph. The number of colors required for a proper edge coloring of G is
greater than or equal to the maximum degree of any vertex of G.
Theorem 2.2.2 (Vizing) [24]. The edge chromatic number of a k-regular graph is either k or k+1.
Theorem 2.2.3. The edge chromatic number of K2n is 2n1.
Theorem 2.2.4. The edge chromatic number of K2n1 is 2n1.

An r-factor of a graph G is an r-regular spanning subgraph of G.


Of course a graph with an odd number of vertices never has a 1-factor. (Why?) But there are also some
connected graphs with an even number of vertices that have no 1-factors.

2.3 Decomposition and Hamilton Cycles


More generally, a graph G is said to be decomposable into subgraphs H 1 , H 2 , . . . , H t if any two subgraphs
Hi and Hj have no edges in common, and the union of all the subgraphs Hi is G.
Notice that all the vertices could appear several times. The union refers to an edge union.
Hamilton cycle
A cycle in G that contains every vertex of G

Theorem 2.3.1 (Lucas) The complete graph K2n+1 has a decomposition into n Hamilton cycles.
Theorem 2.3.2. K2n has a decomposition into n1 Hamilton cycles and a 1-factor.

remark: cant decompose K2n into all HCs because every vertex has odd degree
Theorem 2.3.5. A snark has no Hamilton cycle.

A Hamilton path in a graph G is a path that contains every vertex of G.


Since every path is also a tree, a Hamilton path in G is a spanning tree of G.
Hamilton cycle => Hamilton path
Hamilton path !=> Hamilton cycle
There are graphs that have a Hamilton path but no Hamilton cycle.
A trivial example is the path Pn.
A more interesting example is the Petersen graph
We have seen that K2n has no decomposition into Hamilton cycles. However, it does have a decomposition
into Hamilton paths!

2.4 More Decomposition


Theorem 2.4.1. Let G be a connected graph. Then G is a tree if and only if every edge of G is a bridge.
Theorem 2.4.2. A connected graph with an even number of edges is decomposable into subgraphs
each isomorphic to a path of length two.

3.1 Eulerian Circuits


Walk
a walk in a pseudograph G is an alternating sequence between vertices and edges (not necessarily distinct)

A 1 e1 A 2 e2 A 3 . . . A n1 en1 A n
Trail
A trail in a pseudograph G is a walk in G with the property that no edge is repeated.

Path
A path in a pseudograph G is a trail in G with the property that no vertex is repeated.
A closed trail, also called a circuit, is a trail whose endpoints are the same vertex.
A closed path, usually called a cycle, is a path whose endpoints are the same vertex.
The shortest cycle possible in a graph has length three, and we often call this a triangle.
Euler Circuit
A circuit that contains every edge of G

class
Theorem 3.1.2 (Hierholzer). If G is connected, and the degree of every vertex of G is even, then G has
an Eulerian circuit
Theorem 3.1.4. If a pseudograph G is regular of degree 4, then G has a decomposition into two 2factors.

class
Theorem 3.1.5. A pseudograph G has a decomposition into cycles if and only if every vertex of G has
even degree.
Theorem 3.1.6. A pseudograph G has an Eulerian trail if and only if G is connected and has precisely
two vertices of odd degree.

class
Theorem 3.1.7. (Listing). If G is a connected pseudograph with precisely 2h vertices of odd degree, h
0, then there exist h trails in G such that each edge of G is in exactly one of these trails. Furthermore,
fewer than h trails with this property cannot be found.

3.2
class
Theorem 3.2.1. A regular graph of even degree has no bridge.
Theorem 3.2.3 (Petersen) [18]. A cubic bridgeless graph G has a decomposition into a 1-factor and a 2factor.
Theorem 3.2.4. Every cubic bridgeless graph is decomposable into paths of length three.

class
Theorem 3.2.2. A cubic graph that contains a bridge is not decomposable into three 1-factors.

4.1 A theorem of Turan


largest to mean the graph with the most edges, given the number of vertices.
The number of vertices in a graph G is called the order of G.
The number of edges in G is called the size of G.
An induced subgraph of a graph G is a subgraph of G obtained by taking a subset W of the vertices of G
together with every edge of G that has both endpoints in W

4.2 - Cages
g-cage: a 3 regular graph where the girth is g

K4 is the unique 3 cage all other 3 cage isomorphic


K3,3 the unique 4 cage
Peterson graph the unique 5 cage
Heawood graph unique 5 cage

4.3 Ramsey Theory


Theorem 4.3.1 (Ramsey) [19]. For every number n, there is a number r(n) such that any edge-coloring of
the complete graph with r(n) vertices using red and blue must contain either a red Kn or a blue Kn .
Theorem 4.3.2. For every m and n, there exists the Ramsey number r (m, n) such that any edgecoloring of K r(m,n) with red and blue contains a red K m or a blue Kn . Furthermore, r(m, n) satisfies the
inequality
r(m, n) r(m1, n) + r(m, n1).
r(1, n) = 1

r(2, n) = n

r(3,3) = 6

r(3,4) = 9

r(3,5) = 14

r(3,6) = 18

r(3,7) = 23

r(3,9) = 36

r(4,4) = 18.

5.1 Counting 1-Factors


5.2 Cayleys spanning tree formula
Theorem 5.2.1. The number of spanning trees in Kn is S(K n )

= nn2

prufer sequence

5.3 More spanning trees


S(Km,n ) = mn1 nm1

n1

Theorem 5.3.1. The number of spanning trees in the graph K2,n is n2

##6.1 Magic Graphs


magic labelling
labelling the q edges of a graph with the numbers 1,..,q so that at each vertex, the sum of the labels of the edges
adjacent to the vertex are all the same
anti-magic labelling
labelling the q edges 1,..,q so that the sum at of the labels at each vertex is dierent.

Theorem 6.1.1. Kn,n is magic for n 2.

Using magic square


Theorem 6.1.3. If a graph G is decomposable into two magic spanning subgraphs G1 and G2 where G2
is regular, then G is magic.

6.2 Conservative graphs


directed graph: a graph where each edge is given a direction. Each edge is an ordered pair.
Kirchos current law: says that in a conservative labelling of a directed graph the sum of the labels
entering a vertex equals the sum of the labels of the edges leaving the vertex.
Kircho current law holds and
A graph with q edges that can be so directed and labeled with the numbers 1, 2, 3,, q is called
conservative.
Theorem 6.2.1. If G is decomposable into two Hamilton cycles, then G is conservative.
Theorem 6.2.2 (Kirchhos Global Current Law). If G is a labeled, directed graph such that Kirchhos
Current Law holds at every vertex of G except a particular vertex a, then Kirchhos Current Law also
holds at the vertex a.
Theorem 6.2.3. If G is decomposable into two subgraphs H 1 and H 2 , and if H 1 is conservative, and

H2 is strongly conservative, then G is conservative. Moreover, if both H1 and H2 are strongly


conservative, then G is strongly conservative.

Theorem 6.2.1*. If G is decomposable into two Hamilton cycles, then G is strongly conservative.
Theorem 6.2.4. If G is a graph with n vertices, where n is odd, and G is decomposable into three
Hamilton cycles, then G is strongly conservative.
Theorem 6.2.6. For n 3, the wheel with n spokes, W n , is conservative.
Theorem 6.2.7. If n is even, n 4, then Kn is conservative.

If strongly conservative then also conservative

7.1 Spanning tree algorithm


Spanning tree algorithm
1. Select any edge e with endpoints v1 and v2 set i=1 and j=3
2. Check if the neighbors of vi are in the existing tree
3. If there is a neighbor of vi not in the graph then label this neighbor vj and add the edge vi vj to the tree if j
is equal to the number of vertices in the graph stop.
Otherwise set j=j+1 and i=i+1 and return to Step 2
Kruskal algorithmn
1. Select smallest weight edge add them provided they dont create a cycle
2. Check if you have a spanning tree.

7.2
Know the terms matching, maximum matching, M-alternating path, M-augmenting path. Know how to use
the Hungarian algorithm to find a maximal matching in a bipartite graph.
A matching in a bipartite graph is a subgraph that is regular of degree 1.
That is, a matching is a set of edges with no endpoints in common. A maximum matching is a matching that
has the most edges possible.
If in the bipartite graph under consideration there are equal numbers of red and blue vertices, then a
maximum matching could be a 1-factor. Certainly if the graph contains a 1-factor, then the 1-factor will be a
maximum matching.

8.1 Planar Graphs


A planar graph is called maximal planar if adding an edge between any two nonadjacent vertices results in a
nonplanar graph.
Theorem 8.1.1. (Eulers Polyhedral Formula). If a plane drawing of a connected graph with p vertices
and q edges has r regions, then
p q + r = 2.
Theorem 8.1.2. If G is a maximal planar graph with p vertices and q edges, p 3, then q = 3p6.
Theorem 8.1.3. A planar graph with p vertices, p 3, has no more than 3p6 edges.
Theorem 8.1.4. The graph K5 is not planar.
Theorem 8.1.5. If G is a planar bipartite graph with p vertices and q edges, p 3, then q 2p4
Theorem 8.1.6. The complete bipartite graph K 3,3 is not planar.
Theorem 8.1.8. Suppose G is a maximal planar graph with p vertices and q edges, p 4. Let pi denote
the number of vertices of degree i. Then

3p 3 + 2p 4 + p5 = 12 + p 7 + 2p8 + 3p9 + 4p 1 0+. . . .

8.2 Four color theorem


lune
A plane drawing of a multigraph may contain regions that have only two edges, called lunes.
map
A plane drawing of a connected, bridgeless, planar multigraph is called a map.
normal map
the multigraph is cubic (degree 3 at each vertex) in addition to the above properties, the drawing is called a normal
map.
lune
A plane drawing of a multigraph may contain regions that have only two edges, called lunes.

All planar graphs can be colored by 4 colors


if an edge lies between
countries colored

color the
edge

1 and 2

3 and 4

1 and 3

2 and 4

1 and 4

2 and 3

c.

Theorem 8.2.3 (Tait). If the edges of a normal map can be properly colored by three colors, then the
countries of a map can be colored by four colors.
Thm 8.2.5 In a normal map if edges can be colored by 3 colors then we can color the vertices
black+white so that around any region
num of black vertices - num of white vertices = 0 mod 3

8.3 Five color theorem


Can all planar graphs be colored by five colors?
ofc since more colors means its easier to have have adjacent vertex colors be the same.

8.4 - Graphs and geometry


Theorem 8.4.1 (Wagner). Every planar graph has a plane drawing where every edge is a straight line.

Construct a graph by considering every coin as a vertex and declaring that two vertices are adjacent if and
only if the coins touch. If a graph can be constructed in this way, then it is called a coin graph.
Theorem 8.4.2 (Koebe). Every planar graph is a coin graph.

If all the coins have the same size, the coin graph is called a penny graph.
A region in a plane drawing of a planar graph is called convex if whenever any two points on the boundary of
or in the region are connected by a straight line segment, the straight line segment lies entirely inside the
region.

9.1 Crossing number


Theorem 9.1.2. (Kuratowski). If G is a nonplanar graph, then G contains a subgraph that is a subdivision
of K5 or of K3,3 .

This minimum number is called the crossing number of G, denoted cr(G)


cr (K5) = 1 and cr( K3,3)= 1.
cr(K6) = 3
According to Theorem 9.1.2, if the crossing number is greater than zero, then the graph contains either a
subdivision of K5 or a subdivision of K3,3
cr(G) = 0 if G is planar since crossing number is minimal.
We dont know exactly how much greater >1 crossings. But we know its not planar cr(G)> 1
What is a simple drawing of a graph?

9.2 Thickness and Splitting Number


9.3
10.1 Rotations of graphs
Theorem 10.1.2. Given a connected graph with p vertices and q edges, and a rotation p which induces
r(p) circuits, the inequality
p q + r(p) 2
holds. Furthermore, the alternating sum
p q+r(p)
is even.

How to find # of rotations from scheme?

10.2
10.3

You might also like